Object-token individuation protects targets from object substitution masking

Object Substitution Masking (OSM) occurs when a very sparse mask, such as a group of four single-pixel dots, is presented simultaneously with the target, such as a circle with a small gap that must be localized (Di Lollo, Enns, & Rensink, 2000). Despite the minimal opportunity for contour interference afforded by the sparse masks, when a mask lingers following target offset, severe masking can occur. Phenomenologically, it can appear as though the target is " sucked" out of the array by the mask. Using apparent motion, we showed that OSM can occur even when no mask appears at the target location, as long as a stimulus that is perceived as the same object token as the target lingers elsewhere in the display (Lleras & Moore, 2003). These findings suggest that at least a component of OSM occurs at object-level representations of the scene. In the current study OSM was reduced, and in one case eliminated, through manipulations that facilitated the establishment of separate object tokens for the mask and the target. Apparent motion that depicted the mask as "sliding" into place over the target, presentation of the mask and target in different hues and luminances, configurally grouping the mask with a background of dots away from the target, and finally, grouping the mask away from the target using distinct motion cues all served to reduce OSM. These results further implicate a role of object-level representations in OSM, and provide insight into the source of the interference. Specifically, we suggest that the early establishment of a separate object token for the target protects its representation from being overwritten by the continued sampling of information from a display that includes only the mask. If true, then this form of masking could provide insight into the updating processes that allow for the spatio-temporal coherence of objects, despite changing informational representations. Think: "It's a bird, it's a plane, it's superman!".